$30k challenge to Interpreter’s “Team Bayes”

[malkie]

Kyler has a new screed up:

The TLDR
It seems unlikely that Nephi could have built and sailed a boat from the Arabian Peninsula to the New World.

Though some critics have labeled Nephi’s voyage as an impossibility, those perceptions are largely based on the assumption that Nephi had to have built a Renaissance-style sailing vessel, as if the Nina, Pinta, or the Santa Maria were Nephi’s only options. That assumption doesn’t hold, given a variety of demonstrated oceanic voyages that rely only on ancient technology. Given the success rate of those voyages, and a conservative estimate of 16,000 man-hours of available labor for Nephi and his family, I put the probability of him being to build a boat at p = .3085 and the probability of him being able to sail it the requisite 17,000 miles at p = .0265. The odds would have been stacked against Nephi, but not enough to drastically alter the Book of Mormon’s authenticity.

Evidence Score = -2 (reducing the probability of an authentic Book of Mormon by two orders of magnitude)

That, ladies and gentlemen, is how you turn a mountain into a mole hill.

Perhaps this point has been raised already - if so, forgive me for not having found it in the 150 or so posts that exist on the thread at this time.

Is there good mathematical justification for using probabilities with 3 or more significant figures?

For example, the probability of Nephi's being able to build a boat is set at p = .3085, based on something to do with the abilities of Vikings to build boats, plus a whole lot of assumptions:

To get a probability estimate here, we’ll also need to estimate the variability in boat construction time. Given differences in the design of the ship and expertise of the work involved, that variability is probably pretty large. But let’s assume that ship construction would have a pretty tight standard deviation relative to its mean, say around 1000 hours (the smaller the deviation the better it would be for the critics). That would mean that Lehi’s ship is half a standard deviation away from our estimated mean, and that 30.85% of ships that size would require 16,000 hours or less of labor.

My inclination would have been to make my guess a small fraction, convert that to a decimal, and then round to one sig. fig.

Although I'm no expert, I have taken some math and science courses in the past, and seem to remember that the more uncertainty there is in individual numbers used in a calculation, the fewer significant figures you are entitled to use.

At one point in the current episode he calculates Risk/Mile as .000106762, based on 11 "drift voyages".

Six significant figures? Really?

[drumdude]

I brought it up early on. Kyles degree is not in mathematics, it’s in experimental psychology.

The fact that he doesn’t realize the significance of the number of sig figs that he’s using speaks volumes.

[Physics Guy]

Experimental psychologists use a lot of statistics. The good ones understand and use it well.

Of all fields, though, I think experimental psychology may well be the worst for producing innumerate graduates to whom statistics is a magical dark art. So as a credential for stats expertise, a BA in Psych counts for me like a raise in poker. It could mean a strong hand but it could also be an outrageous bluff.

And yes, positing four- and six-figure probabilities as crudely based estimates shows he is well out of his depth.

[Dr Moore]

At one point in the current episode he calculates Risk/Mile as .000106762, based on 11 "drift voyages".

Six significant figures? Really?

More numbers makes it look more “academic” I guess. It’s all for show and he was pretty clear on this board that he almost just posted the project on a personal blog for he and his friends to have fun with.

[Dr Moore]

Kyler has a new screed up:

The TLDR
It seems unlikely that Nephi could have built and sailed a boat from the Arabian Peninsula to the New World.

Though some critics have labeled Nephi’s voyage as an impossibility, those perceptions are largely based on the assumption that Nephi had to have built a Renaissance-style sailing vessel, as if the Nina, Pinta, or the Santa Maria were Nephi’s only options. That assumption doesn’t hold, given a variety of demonstrated oceanic voyages that rely only on ancient technology. Given the success rate of those voyages, and a conservative estimate of 16,000 man-hours of available labor for Nephi and his family, I put the probability of him being to build a boat at p = .3085 and the probability of him being able to sail it the requisite 17,000 miles at p = .0265. The odds would have been stacked against Nephi, but not enough to drastically alter the Book of Mormon’s authenticity.

Evidence Score = -2 (reducing the probability of an authentic Book of Mormon by two orders of magnitude)

That, ladies and gentlemen, is how you turn a mountain into a mole hill.

The whole project is, by his own admission, little more than a fun blog hobby. Rather like a lengthy attempt to describe a convoluted perpetual motion machine. It’s a waste of time for serious people.

This episode as I read the summary, attempts to treat the transoceanic voyage of Lehi and family. I wonder what John Larsen would have to say about the 1 in 100 odds ascribed to that story.

If Kyler is willing to consider internal Book of Mormon stories as probabilistic evidences, then he has to consider all of them. Or it’s just cherry picking. Which is exactly what it is.

[drumdude]

It's not really worth investing a lot of time in response, but we can easily chain some probabilities on the back of napkin:


Probability of finding enough trees

X

Probability of finding enough sheep for sails

X

Probability of finding an easily accessible vein of ores

X

Probability of building a smelter to smelt all the ore

X

And on and on and on


1:100 is just ridiculous on the face of it.

[Dr Moore]

It's not really worth investing a lot of time in response, but we can easily chain some probabilities on the back of napkin:


Probability of finding enough trees

X

Probability of finding enough sheep for sails

X

Probability of finding an easily accessible vein of ores

X

Probability of building a smelter to smelt all the ore

X

And on and on and on


1:100 is just ridiculous on the face of it.

And not just the Lehite vessel, which by the way had about 7 people to build. But what about the many Jaredite barges as well? If each of those boats also has, per KR, a 2% chance of making the trip, then Kyler’s monkey math would say we are talking about fairly astronomical odds of a whole fleet (12?)of them making the journey with no sail, windows, decks or navigation, with animals and a plug stopper for air and waste?


Also, there is other evidence about these ship journeys. No one died on the journey. Odds of that happening?

[malkie]

I brought it up early on. Kyles degree is not in mathematics, it’s in experimental psychology.

The fact that he doesn’t realize the significance of the number of sig figs that he’s using speaks volumes.

My apologies for bringing it up again.

That you did so makes you a pretty astute fellow, in my opinion

[drumdude]

I brought it up early on. Kyles degree is not in mathematics, it’s in experimental psychology.

The fact that he doesn’t realize the significance of the number of sig figs that he’s using speaks volumes.

My apologies for bringing it up again.

That you did so makes you a pretty astute fellow, in my opinion

Honestly it makes me question my sanity. I can't believe someone who got all the way to the PhD level doesn't realize something we're all taught in high school.

[malkie]

I brought it up early on. Kyles degree is not in mathematics, it’s in experimental psychology.

The fact that he doesn’t realize the significance of the number of sig figs that he’s using speaks volumes.

My apologies for bringing it up again.

That you did so makes you a pretty astute fellow, in my opinion

Honestly it makes me question my sanity. I can't believe someone who got all the way to the PhD level doesn't realize something we're all taught in high school.

I cannot imagine anyone with a PhD in experimental psychology not being very familiar with basic math, and with probability and statistics at a fairly advanced level.